AmirHossein Sojoodi

AmirHossein Sojoodi

PhD student in Computer Engineering at Queen's University

Arm-forge Debugger (Arm DDT)

Published:
Updated:
1 minute read

Arm Forge is a set of tools for HPC development, debugging, and profiling. The server version is not free, but if you have access to a cluster that has Arm Forge as one of its module, then you can access it with its free client.

  1. Load Arm Forge module in the cluster node. If you want a specific version you need to specify it, like:
module load arm-forge/21.0.3

  1. Download and install the same version of Arm Forge client.

  2. In the client, add the remote server configuration. The GUI asks to input any password. If you want to pre-load any modules to be able to use by the application, you can load them in your ~/.bashrc. If you find another way, please share with me.

  3. After the connection is established, you can run and debug the application by browsing the executable file, setting its arguments, passing MPI options, and selecting what you want to profile (CUDA, OpenMP, Memory, …).

Manually launch the backend

With Arm-forge you can launch the backend first and then attach the clients later. For this purpose, follow these steps:

  1. Configure and connect to the remote connection.
  2. Select Manual Launch (Advanced) in the home screen.
  3. Select the number of processes as the number of MPI ranks you would like to attach.
  4. Specify the plugins. There is no need to pre-load any modules or run scripts from Arm-forge.
  5. After starting the backend, run the application with a command like this (options are arbitrary):
mpirun -np 2 /path/to/forge-client/on/server /path/to/ddtsession/file application [args]

# e.g.:

mpirun -np $(np) --display bind --mca pml ucx -x UCX_LOG_LEVEL=trace -x UCX_LOG_FILE=ucx.%p.log \
  /gpfs/fs1/scinet/mist/rhel8/software/2021a/opt/base/arm-forge/22.0.2/bin/forge-client \
  --ddtsessionfile ~/.allinea/session/connection.local-1 \
  application [args]
  1. Simply debug the program

You May Also Enjoy

Utilize PTX Just-In-Time (JIT) Compilation in CUDA

4 minute read

Published:

In this post I’ve written about how to utilize PTX Just-In-Time (JIT) compilation in CUDA. PTX is a low-level assembly-like language that is used to represent the GPU code. The PTX code is then compiled to the machine code by the NVIDIA driver at runtime. This process is called Just-In-Time (JIT) compilation. But before I write about how to use PTX JIT compilation, I’ll provide some background on why you might want to use it.

Collection of my dotfiles (public version)

1 minute read

Published:

This is a collection of my dotfiles that I use on my Linux system. I have written about them in various posts, but I thought I can put them all in one place. You can checkout the github repository here, too.

Tmux Tips and Tricks

10 minute read

Published:

(If you are already familiar with tmux, and you are looking for a simple .tmux.conf, go to the end of this post.)

Setup LAMMPS

4 minute read

Published:

Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) is a classical molecular dynamics code that can be used to model atoms or, more generally, as a parallel particle simulator at various scales. The complete documentation of LAMMPS can be found here. In this post, I will provide a guide on how to setup LAMMPS on a Linux machine. My setup is on a cluster with NVIDIA GPUs, UCX, and OpenMPI. Also, we have a built-in module system to load the necessary modules.